CoolFace
Modelpublic

metafloor-ai/manifest-orchestrator-0.8b-v0.6.0

sourceHugging Facecc-by-nc-4.0updated 1mo agoView on Hugging Face
0likes10downloads
Model Card

Manifest 0.8B · v0.6.0

The lightweight sprinter of the Manifest suite — instant supply-chain answers at the lowest cost, small enough to run almost anywhere.

Manifest is MetaFloor's suite of supply-chain expert models — purpose-built specialists in procurement, demand planning, warehouse operations, supplier relationship management, risk & resilience, transportation, inventory, and order fulfilment.

  • —Family: Manifest · This model: Manifest 0.8B (0.8-billion-parameter base)
  • —Tier: entry — speed and cost first
### Preferred 79.9% of the time over the base model On 134 held-out expert questions, an independent LLM judge panel picked this model's answer over the untuned base model's answer 79.9% of the time (95% CI 73.1–86.2%). With both models given the same answer format, it is still preferred 85.4% of the time — the gain is real domain knowledge, not just presentation.

What's new in v0.6

  • —Retrained on MetaFloor's expanded ~32k-example supply-chain dataset (up from ~12.5k in v0.5).
  • —The evaluation benchmark grew to 134 held-out questions (from 116) — so v0.6 headline figures are measured on a larger, harder set than the v0.5 cards.
  • —A new [Manifest 9B](https://huggingface.co/metafloor-ai/manifest-orchestrator-9b-v0.6.0) tier is joining the family (coming soon); the 35B orchestrator is retired from this release.
  • —This is a new versioned repo — the previous v0.5 release (`metafloor-ai/manifest-orchestrator-0.8b`, unchanged) keeps its own history and download stats.

v0.5 → v0.6, same 134-question benchmark (preferred over the untuned base):

ReleasePreferred over basevs base + matched format
v0.5.057.3%68.5%
v0.6.079.9%85.4%

(The baseline throughout is the untuned `Qwen/Qwen3.5-0.8B` — every win-rate above is measured against it.)

Best for

Speed and volume: quick, cheap, everywhere.

  • —High-volume Q&A and front-line operational lookups
  • —In-app assistants and chat where latency and cost matter
  • —Edge / on-device / CPU-only deployment (adapter is just ~44 MB)
  • —Fast first-draft answers you refine or route onward

(Need more depth? Step up to Manifest 2B or Manifest 4B below — same suite, same prompt format.)

See the difference

Same question. Base model vs Manifest 0.8B.

Ask (COO of a $5.8B enterprise, mostly manual): "Explain the core concept of MEIO and why it's relevant given our manual, single-echelon approach." Base model → "### Core Concept of MEIO. Multi-Echelon Inventory Optimization is a strategic supply chain management technique that shifts inventory from a single-echelon (one-way) model to a multi-echelon (two-way) model…" — a generic, slightly muddled textbook definition. Manifest 0.8B → "MEIO optimizes inventory across several levels of the supply chain — the few strategic nodes and the many local ones — so that total cost is minimized while still meeting service targets…" — crisp and on-point.

The Manifest family

Two kinds of models:

🧭 Orchestrators — general-purpose, handle any supply-chain area

ModelSizePreferred over baseStatus
Manifest 0.8B0.8B79.9%✅ available
Manifest 2B2B85.1%✅ available
Manifest 4B4B95.1%✅ available
Manifest 9B9B—🔜 coming soon

🎯 Domain-experts — specialized for a single area

ModelPreferred over baseStatus
Manifest Specialist · Risk & Resilience100% (20 / 0)✅ available
Manifest Specialist · Inventory Optimization80.0%✅ available
Manifest Specialist · Demand Planning82.5%✅ available

Orchestrators are scored on the general supply-chain benchmark; domain-experts on their focused domain benchmark (20 items each — treat those exact figures as directional).

Not just a model — a full stack

Manifest is backed by everything needed to build it and trust it:

  • —A purpose-built dataset — thousands of supply-chain instruction–response pairs spanning 8 sub-domains and every company scale, generated by a seed-driven operator-as-teacher pipeline.
  • —A reproducible training pipeline — documented LoRA fine-tuning.
  • —An independent benchmark — 134 held-out expert questions, scored blind by a panel of LLM judges.

We built the model, the data, and the evaluation.

How to use

Manifest 0.8B is a LoRA adapter (~44 MB), applied on top of its base model at load time.

python
from peft import PeftModel
from transformers import AutoModelForCausalLM, AutoTokenizer

base = "Qwen/Qwen3.5-0.8B"  # base model — see "Built on" below
tok = AutoTokenizer.from_pretrained(base)
model = AutoModelForCausalLM.from_pretrained(base, device_map="auto")
model = PeftModel.from_pretrained(model, "metafloor-ai/manifest-orchestrator-0.8b-v0.6.0")

SYSTEM = "You are a senior supply chain expert. Answer correctly and concisely."
user = (
    "I'm an inventory planner at a ~$8M small business: ~11k active SKUs, 4 suppliers, "
    "2 network nodes, ~164-day avg lead time. How should I set safety stock as I move off spreadsheets?"
)
msgs = [{"role": "system", "content": SYSTEM}, {"role": "user", "content": user}]
inputs = tok.apply_chat_template(msgs, add_generation_prompt=True, return_dict=True, return_tensors="pt")
out = model.generate(**inputs, max_new_tokens=512, temperature=0.7)
print(tok.decode(out[0][inputs["input_ids"].shape[-1]:], skip_special_tokens=True))

Prompt tip: Manifest is trained to condition on the scenario — include the asker's role and operating scale (revenue, SKUs, suppliers, nodes, lead time) in the message for the sharpest, most tailored answers.

How it was measured

134 held-out expert questions across 8 supply-chain areas. Each question is answered by Manifest and by the base model (given the same answer format); an independent two-model LLM judge panel then picks the better answer. Manifest 0.8B was preferred 85.4% of the time (95% CI 79.5–90.7%; 111 wins / 16 losses / 7 ties over 134). Benchmark: [supply-chain-eval](https://huggingface.co/datasets/metafloor-ai/supply-chain-eval).

Training details

MethodLoRA (PEFT 0.20.0), rank 16 / alpha 16 / dropout 0.05
Target modulesall attention + MLP projections
Trainable params6,389,760 (~1.25% of the 0.51B base)
Epochs3
Training examples~32,000
Final loss1.71 (from 3.04)

Training data: supply-chain instruction–response pairs from the seed-driven operator-as-teacher pipeline — a deterministic engine emits a unique seed per example (area, sub-area, persona, question type, realistic numeric scenario) and a strong teacher model writes the matching answer. The training data is drawn from MetaFloor's proprietary ~32k-example supply-chain dataset, which is not open-sourced — only the held-out evaluation benchmark (supply-chain-eval) is public.

Intended use & limitations

  • —Intended use: low-latency, low-cost decision-support and drafting for supply-chain professionals.
  • —Out of scope: not legally binding, contractual, or safety-critical guidance; no access to your live systems or real-time data. Verify outputs before acting on them.
  • —Limitations: English-only; trained on synthetic (model-authored) data; as the smallest model in the suite it trades some depth for speed and cost; standard LLM risks (hallucination, outdated facts) apply.

License

Manifest models and the supply-chain-eval benchmark are released under CC-BY-NC-4.0 — free for research and non-commercial use, with attribution. Commercial use requires a license from MetaFloor — get in touch at metafloor.ai.

Built on

Manifest 0.8B is a LoRA adapter over Qwen/Qwen3.5-0.8B (used under its own license); the base model is required to load the adapter.

Citation

bibtex
@misc{metafloor_manifest_0_8b,
  title  = {Manifest 0.8B: a supply-chain expert model (MetaFloor Manifest suite)},
  author = {MetaFloor AI},
  year   = {2026},
  howpublished = {\url{https://huggingface.co/metafloor-ai/manifest-orchestrator-0.8b-v0.6.0}}
}

<!-- manifest-suite:order -->